Dataset with English words classified along CEFR categories, tokenized forms based on sentencepiece tokenizer.